This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: gcc halfword bug


Jeff, 

  Can this go into egcs-1.1 please.  It fixes a bug with handling HImode 
when the compiler tries to reload a memory-memory move.

Richard.

Mon Aug 17 11:25:52 1998  Richard Earnshaw <rearnsha@arm.com>

	* arm.h (SECONDARY_INPUT_RELOAD_CLASS): Return NO_REGS if compiling
	for architecture v4.

Index: arm.h
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/config/arm/arm.h,v
retrieving revision 1.10
diff -p -r1.10 arm.h
*** arm.h	1998/06/26 07:13:25	1.10
--- arm.h	1998/08/17 10:13:41
*************** enum reg_class
*** 874,880 ****
  
  /* If we need to load shorts byte-at-a-time, then we need a scratch. */
  #define SECONDARY_INPUT_RELOAD_CLASS(CLASS,MODE,X)		\
!   (((MODE) == HImode && TARGET_SHORT_BY_BYTES			\
      && (GET_CODE (X) == MEM					\
  	|| ((GET_CODE (X) == REG || GET_CODE (X) == SUBREG)	\
  	    && true_regnum (X) == -1)))				\
--- 874,880 ----
  
  /* If we need to load shorts byte-at-a-time, then we need a scratch. */
  #define SECONDARY_INPUT_RELOAD_CLASS(CLASS,MODE,X)		\
!   (((MODE) == HImode && ! arm_arch4 && TARGET_SHORT_BY_BYTES	\
      && (GET_CODE (X) == MEM					\
  	|| ((GET_CODE (X) == REG || GET_CODE (X) == SUBREG)	\
  	    && true_regnum (X) == -1)))				\

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]